More On Friedman's "Positive Economics"

In October, I linked to an interesting article about the philosophy and methodology about economics that defended the Austrian view and attacked Milton Friedman's positivist view. It contained in short two key arguments:

1) Concepts are abstractions of certain aspect on groups of thing that omit other characteristics without denying them. For example an eye is an organ that detect light and convert it into electro-chemical impulses in neurons. Eyes function and look differently depending on which species and to a lesser extent even which individual they are on, but the concept of eye only contain the essential characteristic of detecting light and converting it into electro-chemical impulses in neurons. These other characteristics aren't included but not denied either. These other characteristics exist in some form but may come in any form.

The implication of this is that Friedman was wrong when he asserted in his essay "The Methodology of Positive Economics" that a completely realistic theory of markets would have to include for example the different eye- and haircolors of traders. His point with that is that since this according to him shows that economic theory has to be unrealistic, there's no harm in using false assumptions. But economic analysis is based on the economic concepts used that only include relevant characteristics of the markets, and the different eye- and haircolors of traders is usually not among them. That doesn't mean that the existence of different eye- and haircolors is denied, but it is omitted because it is usually completely irrelevant. Omitting facts irrelevant for the analysis isn't unrealistic, unlike the use of directly false assumptions, like for example "perfect competition".

2) Economic analysis isn't so much about how one action or event causes another later in time, but about how actions and events causes things to be different compared to a counter-factual scenario, The relevance of this distinction can be illustrated by looking at currency market movements. QE2 has no doubt caused the U.S. dollar to have a weaker exchange rate than it otherwise would have had, but other factors, primarily the European debt crisis have still caused it to rise in value against many currencies. An analysis that simply involved looking at how things change over time couldn't make that statement about the effects of QE2 because that's not how things have turned out, but a counterfactual analysis can still say that the dollar would have risen even more in value against these other currencies if it hadn't been for QE2.

Similarly, to use Friedman's example of the minimum wage, counterfactual analysis can establish that unless it is so low that it is below the marginal productivity of everyone who have a job or would have found a job, a minimum wage will increase unemployment, but in reality higher minimum wages might be followed by lower unemployment if there are other factors pushing it down.

The point with this is both that this makes it very difficult to really test theories and that counterfactual analysis of economic action makes it possible to establish economic laws depite the fact that humans have free will.

All of this is good, but after re-reading Friedman's essay (or more specifically the 34 out of 42 pages available on Google books), I noticed that Friedman used more arguments that should be adressed.

One, that could be used as an argument against point 1) above is that we can't know which aspects are relevant so the only way of knowing what is relevant is what theory of relevant aspects makes the best predictions. But while it is true that we by some means must investigate the specific characteristics of a certain situation to determine what the specific causal factors are, analysing the predictions of certain possible causal factors isn't necessarily the best way of getting to the truth because as mentioned above there are often different causal factors that work in opposite directions. Furthermore, this question isn't about the correctness of theories but about the applicability of theories. There is a big difference between a theory being incorrect, like the theory that lightning is created by the Norse god Thor, and a theory being inapplicable, like the correct theory of lightning is in places where lightning doesn't occur.

Another is that people and things often behave or move as if a false theory was true. He uses examples like how many objects drop at the speed it would have dropped in vacuum, and that leaves are positioned as if they deliberately tried to maximize the amount of sunlight it receives or that expert billiard players use mathematical formulas to determine how he should make his shots. We all know these explanations aren't true but our observations are still consistent with them. Friedman's point is therefore that it is only natural to use false assumptions in economics, such as the neoclassical mathematical models that says that businessman make their business decisions and household their purchase decisions using advanced mathematics since supposedly they are consistent with predictions about the behavior of business executives and housheholds.

But this conclusion doesn't follow even assuming for the sake of the argument that these models really make correct predictions (a very questionale assumption to say the least). The fact that some false theories make predictions that turn out to be correct, or at least appears to be correct, is the reason why many false theories in for example physics have been able to live on for so long. Most physical phenonenoms that we perceive are entirely consistent with Newtonian physics, even though we now know it is wrong in at least some aspects.

But the difference between physics and economics is that we already know what is correct whereas in physics all theories are ultimately derived from experiments. Thus, while it is correct in physics to base theories on which theory currently is most consistent with predictions from experiment, there is no need to assume false theories in economics just because they make predictions that turn out to be correct. Basing economic theories on predictions therefore brings us away from the truth, unlike in physics where they bring us closer to the truth. Friedman's argument is akin to saying that because crutches facilitates movement for people with wounded legs, people with good legs should also move using crutches even though it impairs movement for them.